RESEARCH Open Access

Geographical distribution and relative riskof Anjozorobe virus (Thailandorthohantavirus) infection in black rats(Rattus rattus) in MadagascarVololoniaina Raharinosy1,2, Marie-Marie Olive1, Fehivola Mandanirina Andriamiarimanana3,Soa Fy Andriamandimby1, Jean-Pierre Ravalohery1, Seta Andriamamonjy1, Claudia Filippone1,Danielle Aurore Doll Rakoto4, Sandra Telfer5† and Jean-Michel Heraud1*†

Abstract

Background: Hantavirus infection is a zoonotic disease that is associated with hemorrhagic fever with renalsyndrome and cardiopulmonary syndrome in human. Anjozorobe virus, a representative virus of Thailandorthohantavirus (THAIV), was recently discovered from rodents in Anjozorobe-Angavo forest in Madagascar. Toassess the circulation of hantavirus at the national level, we carried out a survey of small terrestrial mammals fromrepresentative regions of the island and identified environmental factors associated with hantavirus infection. As wewere ultimately interested in the potential for human exposure, we focused our research in the peridomestic area.

Methods: Sampling was achieved in twenty districts of Madagascar, with a rural and urban zone in each district.Animals were trapped from a range of habitats and examined for hantavirus RNA by nested RT-PCR. We alsoinvestigated the relationship between hantavirus infection probability in rats and possible risk factors by usingGeneralized Linear Mixed Models.

Results: Overall, 1242 specimens from seven species were collected (Rattus rattus, Rattus norvegicus, Mus musculus,Suncus murinus, Setifer setosus, Tenrec ecaudatus, Hemicentetes semispinosus). Overall, 12.4% (111/897) of Rattus rattusand 1.6% (2/125) of Mus musculus were tested positive for THAIV. Rats captured within houses were less likely to beinfected than rats captured in other habitats, whilst rats from sites characterized by high precipitation and relativelylow seasonality were more likely to be infected than those from other areas. Older animals were more likely to beinfected, with infection probability showing a strong increase with weight.

Conclusions: We report widespread distribution of THAIV in the peridomestic rats of Madagascar, with highestprevalence for those living in humid areas. Although the potential risk of infection to human may also bewidespread, our results provide a first indication of specific zone with high transmission. Gathered data will behelpful to implement policies for control and prevention of human risk infection.

Keywords: Hantavirus, Anjozorobe virus, Thailand orthohantavirus, Rodent, small terrestrial mammals, Risk factors,Madagascar, Africa

* Correspondence: jmheraud@pasteur.mg†Equal contributors1Virology Unit, Institute Pasteur de Madagascar, Ambatofotsikely, BP 1274Antananarivo, MadagascarFull list of author information is available at the end of the article

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Raharinosy et al. Virology Journal (2018) 15:83 https://doi.org/10.1186/s12985-018-0992-9

BackgroundHantaviruses are zoonotic viruses whose main reservoirhosts are rodents, although these viruses have recentlybeen detected in insectivores and bats [1, 2]. Human in-fections appear in several areas of the globe. Clinicalsymptoms and severity of disease vary between regionsaccording to the species of hantavirus involved. Two dis-eases associated with hantavirus infections are described,the Hantavirus Cardiopulmonary Syndrome (HCPS) andthe Hemorrhagic Fever with Renal Syndrome (HFSR).Nephropathia Epidemica (NE) is a mild form of HFSR.While NE disease appears to be mild with 0.1% of mor-tality rates, HCPS and HFRS, are more severe with amortality that can reach up to 50 and 15% respectively[3]. HCPS are mainly found in the Americas, and HFRSand NE appears in the Old World [4, 5].Hantaviruses are generally thought to be relatively

host-specific [6], for example, Seoul orthohantavirus(SEOV) is associated with the rat, Rattus norvegicus, [7]and Puumala orthohantavirus (PUUV) is preferentiallyhosted by the bank vole, Clethrionomys glareolus [8].However, studies have shown that spillover infection canoccur between sympatric host species [9, 10]. For ex-ample, the common vole, Microtus arvalis, is the prefer-ential host of Tula orthohantavirus (TULV), but thevirus is also detected in other vole species such asMicrotus agrestis and Arvicola spec. [11]. Hantavirusescan be found in saliva, urine and feces of infected ani-mals and transmission between reservoirs is thought tobe through bites, scratches or inhalation of contami-nated aerosols [12, 13]. Humans usually contract hanta-virus infection by direct contact with rodents or byinhalation of contaminated aerosols from excretion orsecretion from infected animals [14]. Transmissioninter-human is very uncommon, although it has been re-ported for some lineages of Andes orthohantavirus(ANDV) [15, 16].Although relatively few hantavirus studies have been

carried out in Africa and the surrounding regions, since2006 eight hantaviruses have been [14] identified in ro-dents, insectivores and bats in mainland Africa [17–24]whilst recently two hantaviruses were detected in the In-dian Ocean region [25, 26]. The two hantaviruses foundin the Indian Ocean region both belonged to Thailandorthohantavirus (THAIV) specie which was primarilyfound in Southeast Asia [27]. Mayotte virus (MAYOV)harbored by Rattus rattus in Mayotte island [26] andAnjozorobe virus (ANJOV) found in Rattus rattus andthe endemic rodent Eliurus majori in Madagascar [25].Studies have also documented exposure of human

populations in Africa to hantavirus [28–30], and recentlyHeinemann et al. have reported the presence of Bowéorthohantavirus associated with human disease in an Ivor-ian patient [31]. Moreover, as Thailand orthohantavirus is

associated with disease [32], it is probable that hantavirusinfection in Madagascar is also symptomatic. Neverthe-less, the lack of studies in Madagascar prevents us fromhighlighting the potential threat that this virus presentsfor public health and from raising awareness amongst cli-nicians of clinical symptoms associated with hantavirusinfection.In areas of the globe where hantavirus infections in

humans are better monitored, there is strong spatial andtemporal variation in disease incidence and studies havedemonstrated effects of environmental variables, includ-ing climate, habitat and land-use [6]. Climatic conditionsmay influence persistence of hantavirus in the environ-ment and, therefore, transmission rates to humans [33].However, these environmental effects on incidence arethought to be primarily mediated by the impacts on theabundance of infected reservoirs [34, 35]. For individualhantavirus, spatial variation in risk is likely to largely de-pend on the bioclimatic and habitat preferences of theprinciple host, whilst temporal variation in risk may bedriven by inter-annual climatic driven changes in foodavailability and breeding rates [36]. Despite the tendencyfor close associations between individual mammal spe-cies and individual hantaviruses, several studies have alsosuggested that increased mammal diversity may reducethe prevalence of hantavirus infection in the principalhost [37–40], consistent with a “dilution effect”.For areas where data on infection in humans may be

under-recorded, examining the spatial distribution of in-fected reservoirs provides an important first-step in un-derstanding the potential risk to humans, as well asallowing analyses to explore epidemiological processes inthe reservoir populations by identifying environmentalrisk factors. The previous study in Madagascar that de-tected Anjozorobe virus was limited to one forest loca-tion, Anjozorobe-Angavo forest [25]. Here we sampleterrestrial peridomestic small mammals trapped in urbanand rural sites of representative regions of Madagascarto (i) establish the distribution of Anjozorobe virus and(ii) evaluate to what extent spatial variation in infectionrates is related to environmental and host-related vari-ables. Our results will be helpful to implement policiesfor future awareness and surveillance programs for hu-man communities, as well as increasing our understand-ing of the epidemiology of hantavirus in a peridomesticAfrican context.

MethodsSpecimen collectionSampling was carried out in twenty districts ofMadagascar, with a rural and urban zone in each districtexcept for one district where only the rural zone was sam-pled. The urban zone was centered on a health center,whilst the rural zone was located within a randomly

Raharinosy et al. Virology Journal (2018) 15:83 Page 2 of 11

selected commune within 15 km of the health center.Within the urban zone, three sub-areas for trapping wererandomly selected within an 800 m*800 m square, whilstthe rural zone centered on a single village. Terrestrialsmall mammals were captured within a range of habitats.At each zone 15–20 houses were sampled, using onewire-mesh trap (BTS) and one Sherman trap placed insideeach house. Where possible a further BTS trap was set inthe immediate vicinity of the house. Additional BTS trapswere set in trapping lines close to vegetation in (i) areasaround houses (either adjacent to paths or areas for dis-posing of waste) and (ii) around rice fields or other lowground primarily used for agriculture. In urban zones,where possible, additional BTS traps were set in marketsand around the abattoir. Traps were baited with onion, dryfish and carrot and checked each morning for 3 days. Alltrapping was conducted February 2012–April 2012 or Oc-tober 2012–May 2013. Species, gender and weight of eachcaptured animal were recorded. Animals were euthanizedby cervical dislocation and organs were collected and storedin liquid nitrogen before laboratory storage at − 80 °C.

RNA extraction and nested RT-PCRApproximately 100 mg of liver and spleen tissues fromeach individual rodent were ground using Tissuelyser II(Qiagen) with a 5 mm stainless steel beads. Organ super-natants were recovered at 1:10 dilution of culture mediumcontaining 40% of fetal bovine serum and antibiotics.Total RNA was extracted using the TRIzol LS reagents(Invitrogen, Carlsbad, CA) according to the manufac-turer’s instructions. cDNA was prepared using PRO-MEGA Kit One Step PCR and amplified by a previouslypublished nested-PCR protocol that targets conserved re-gion of the L gene [18]. All PCR products were sequenced(Cogenics, Essex, UK).

Statistical analysisWe investigated the relationship between hantavirus in-fection probability in rats and possible risk factors byusing binomial Generalized Linear Mixed Models(GLMM) with a logit link. Individual infection statuswas the response. Explanatory covariates were selectedfor consideration based on their potential to influencetransmission of hantavirus within small mammal popu-lations and included individual-level variables and site-level variables. Given that rodents trapped at the samesite were non-independent, site was included as a ran-dom effect. Continuous covariates were centred andstandardized. Non-linear relationships were consideredby including the square of the variable. To limit thenumber of models being considered we performed stat-istical analysis in two stages: first examining individual-level variables and then site level variables. Model selec-tion was based on likelihood ratio tests (LRT) to

compare nested models, and Akaike Information Criteria(AIC) to compare non-nested models. Models with anAIC within 2 of the model with the lowest AIC (ΔAIC< 2)are equally likely to be the best model [41].The individual-level stage of the analysis considered

weight (as a proxy for age), sex and habitat. Two differ-ent habitat related variables were considered. The firstwas a six-level factor including inside house, outsidehouse, abattoir, market, exterior trap lines and low-ground trap lines. The second was a two-level factor thatcompared inside house rats to rats from all other habi-tats mentioned above. Explanatory variables were firstconsidered in univariate models. However, initial ana-lyses indicated that weight had a very strong effect andmodels without weight sometimes had problems conver-ging, weight was therefore included in all models.The site-level stage of the analysis included all vari-

ables significant from the first stage of the analysis andadded site-level variables individually. A range of vari-ables related to climate were considered. Bioclimate wasclassified into four zones: dry, subarid, subhumid andhumid. The following climatic variables were also ob-tained from WorldClim (http://worldclim.org/version2):Annual Mean Temperature BIO1, Mean Diurnal RangeBIO2, Temperature Seasonality BIO4, Annual Precipita-tion BIO12, Precipitation Seasonality BIO15 and Precipi-tation of Driest Quarter BIO17. WorldClim variableswere extracted from a 10 km radius from a point mid-way between the urban and rural sites. As these climaticvariables were correlated, it would not be possible toconsider them together in subsequent multivariateGLMMs. We therefore also used principle componentsanalysis (PCA) to summarize climatic variation betweensites and considered the first two principal componentsas covariates in the GLMM. In addition to climatic vari-ables, we also examined site type (rural vs urban); seasonwhich was defined as dry season October to Decemberand rainy season January to April; inside house R. rattusabundance (number rats trapped in houses/number oftrap nights in houses), outside house R. rattus abun-dance (number rats trapped outside/number of trapnights outside); host species diversity based on the Shan-non diversity index and host species diversity based onEvenness [42].In each stage, following individual assessment of vari-

ables, non-correlated variables with a p-value for the LRTof < 0.20 were considered for inclusion in a multivariatemodel. Statistical analyses were conducted in R softwareversion 3.3.1 using the following packages: vegan, ade4and lme4 [43–46].

ResultsSeven species of small mammal were trapped. R. rattuswere trapped in nearly all (97%) of the sites (38/39),

Raharinosy et al. Virology Journal (2018) 15:83 Page 3 of 11

whilst R. norvegicus were primarily trapped in the largecities (31% of sites, 12/39). We tested specimens from1242 mammals belonging to seven species: 897 blackrats R. rattus, 125 house mouse Mus musculus, 124Norway rats R. norvegicus and 96 insectivores (Suncusmurinus, Setifer setosus, Tenrec ecaudatus, Hemicentetessemispinosus) (Table 1). The number of animals testedper site ranged from 11 to 62.Overall 113 (9%) small mammals were PCR positive

for hantaviruses, of which 111 were R. rattus and two M.musculus. It is to be noted that all PCR-positive sampleswere confirmed by sequencing. Hantavirus RNA was notdetected in samples from R. norvegicus or insectivores.Genetic analysis of the partial the coding region of the Lsegment (301 nt) of rodent-borne hantaviruses detectedin Madagascar show that these viruses cluster with pre-vious hantavirus detected in Madagascar; Anjozorobevirus (Fig. 1).Positive individuals showed a widespread geographical

distribution, with infected individuals in 90% of districts(18/20) (Fig. 2) and 69% of sites (27/39). The prevalenceof infected R. rattus at district level ranged from 0 to29% (Table 1). The highest prevalences were obtained inthe districts of Moramanga (29%) Fianarantsoa (27%),

Toamasina (26%), Maevatanana (25%), Mandritsara (25%),Farafangana (24%) and Antananarivo-Atsimondrano (23%).At the site level, prevalence in R. rattus ranged from 0 to

80%, with highest prevalences in Moramanga Urban (80%,n = 5), Mandritsara Rural (58%, n = 19), Fianarantsoa Urban(43%, n = 30) and Anjozorobe Urban (43%, n = 7). For the19 sites where both urban and rural sites were tested therewas no correlation between the prevalence in the urbanand the rural site (Spearman rank, p = 0.9).

Risk factors of hantavirus infection in R. rattusFor these analyses, only R. rattus were considered due tothe small number of hantavirus positive from the otherspecies.In the individual-level stage, univariate logistic ana-

lyses indicated that weight was an important factor witha highly significant positive linear effect on hantavirusinfection probability (Table 2). There was no differencebetween males and females. However, significant differ-ences in infection probabilities were observed betweenhabitats. Based on Akaike information criterion (AIC),the model with habitat as a two-level factor appears tosufficiently describe this variation, with houses having alower probability of infection than other habitats

Table 1 Prevalence of hantavirus nucleic acids detected in small terrestrial mammals captured in Madagascar

District n/N of captured species (%) n/N (%)

Rr Rn Mm Sm Ss Hs Te

Antsirabe 3/61 (4.9) 0/9 (0.0) 0/7 (0.0) 0/2 (0.0) – – – 3/79 (2.7)

Miandrivazo 9/70 (12.9) – 0/7 (0.0) 0/3 (0.0) – – – 9/80 (8.0)

Ihosy 1/58 (1.7) – 0/12 (0.0) – 0/2 (0.0) – – 1/72 (0.9)

Nosy-Be 3/24 (12.5) 0/17 (0.0) 0/1 (0.0) 0/13 (0.0) – – – 3/55 (2.7)

Antsohihy 4/34 (11.8) 0/6 (0.0) 0/3 (0.0) 0/11 (0.0) 0/2 (0.0) – – 4/56 (3.5)

Toamasina 6/23 (26.1) 0/12 (0.0) 0/2 (0.0) 0/10 (0.0) 0/3 (0.0) – 0/1 (0.0) 6/51 (5.3)

Sambava 5/86 (5.8) – – 0/3 (0.0) – – – 5/89 (4.4)

Mandritsara 15/61 (24.6) – 0/5 (0.0) – – – – 15/66 (13.3)

Morombe 3/57 (5.3) – 2/3 (66.7) – – – – 5/60 (4.4)

Morondava 1/17 (5.9) 0/29 (0.0) 0/5 (0.0) 0/2 (0.0) – – – 1/53 (0.9)

Belo sy Tsiribihina 0/31 (0.0) 0/11 (0.0) 0/10 (0.0) 0/4 (0.0) 0/4 (0.0) – – 0/60 (0.0)

Fianarantsoa 16/60 (26.7) – 0/9 (0.0) 0/1 (0.0) – 0/1 (0.0) – 16/71 (14.2)

Farafangana 13/55 (23.6) – 0/2 (0.0) 0/8 (0.0) – – 0/2 (0.0) 13/67 (11.5)

Mananjary 5/53 (9.4) – 0/3 (0.0) 0/6 (0.0) 0/4 (0.0) – – 5/66 (4,4)

Antananarivo- Atsimondrano 7/30 (23.3) – 0/10 (0.0) – – – – 7/40 (6,2)

Maevatanana 11/44 (25.0) 0/1 (0.0) 0/2 (0.0) 0/8 (0.0) – – – 11/55 (9,7)

Moramanga 4/14 (28.6) 0/24 (0.0) 0/11 (0.0) 0/1 (0.0) – – – 4/50 (3,5)

Ambovombe-Androy 0/57 (0.0) – 0/16 (0.0) – – – – 0/73 (0.0)

Taolagnaro 1/27 (3.7) 0/9 (0.0) 0/4 (0.0) 0/2 (0.0) – – – 1/42(0.88)

Anjozorobe 4/35 (11.4) 0/6 (0.0) 0/13 (0.0) 0/2 (0.0) – 0/1 (0.0) – 4/57(3.54)

TOTAL 111/897 (12.4) 0/124 (0.0) 2/125 (1.6) 0/76 (0.0) 0/15 (0.0) 0/2 (0.0) 0/3 (0.0) 113/1242 (9.1)

Rr Rattus rattus, Rn Rattus norvegicus, Mm Mus musculus, Sm Suncus murinus, Ss Setifer setosus, Hs Hemicentetes semispinosus, Te Tenrec ecaudatus, N total numberof samples tested, n number of hantavirus positive samples, % hantavirus prevalence

Raharinosy et al. Virology Journal (2018) 15:83 Page 4 of 11

(Table 2). Prevalence of infected rodents living withinhouses or outside were 7.5% (34/456) and 17.5% (77/441)respectively.Principal Components Analysis (PCA) results showed

that the two principal components F1 and F2 represented88% of the total climatic data information (Fig. 3). F1 waspositively correlated with annual precipitation and precipi-tation of driest quarter and negatively correlated withmean diurnal range in temperature and precipitationseasonality. Thus, humid sites were characterized by highvalues of F1, whilst subarid and some subhumid sites thatshowed more seasonal variation had low values of F1. F2exhibited positive correlation with mean annualtemperature (Fig. 3), with dry sites having the highestvalues for F2.The GLMMs in the site-level stage showed strong evi-

dence of climatic effects. In models that included thebest model from the individual-level stage and singlesite-level covariates, the mean diurnal ranges in

temperature, annual precipitation and F1 from the PCAanalysis were all associated with variation in infectionprobability based on LRT (Table 3). There was evidenceof a non-linear relationship with F1 and some suggestion ofa non-linear effect for mean diurnal range in temperatureand annual precipitation (Table 3). Infection probabilitiesincreased with increasing precipitation (OR = 2) and in-creasing values of F1 (at least to some kind of plateau) anddecreased with an increasing diurnal range in temperature(OR = 0.56). There was no evidence of any differencebetween rural and urban sites. There was also no effect ofseason, rodent abundance measures or host diversity mea-sures. Consequently, as the climate measures were highlycorrelated no further multi-variate models were considered.Comparison of AIC values from the univariate models

indicated that models including precipitation performedbest (Table 3). Although, the model with the lowest AICincluded a non-linear effect of precipitation, a simplelinear effect was similar in its ability to describe the data.

Fig. 1 Phylogenetic tree of hantavirus based on the partial sequences of the L segment coding region. Phylogenetic tree generated by the Neighbour-joining methods and Kimura-2 parameter, based on the alignment of the coding region of the L partial segment 301 nucleotides long of rodents-bornehantaviruses detected in Madagascar. ●Reference Sequences ■ Outgroup sequences obtained from Genbank (http://www.ncbi.nlm.nih.gov/nuccore/. Ourstrains are indicated by Z and the initial of the district followed by the name of the isolation site. In bold are the sequences of hantavirus detected in Musmusculus. Only bootstrap percentages ≥70% (from 1000 resampling) is indicated. Scale bar indicates nucleotide substitution per site

Raharinosy et al. Virology Journal (2018) 15:83 Page 5 of 11

The model with a non-linear effect of F1 was marginallyworse at explaining the data (ΔAIC = 2.32). Thus, highprecipitation appears to be the principle climatic driverof high hantavirus infection probability.

DiscussionOur national-scale study demonstrates that THAIV iswidespread in the peridomestic small mammal commu-nities of Madagascar, primarily infecting R. rattus, themost common species in our study, in both rural andurban sites. Rats living outside houses are more likely tobe infected than those inhabiting houses and hantavirus

prevalence is higher in humid areas of the island. Weightalso influenced infection probability, as several studieshave shown before for this chronic infection [47, 48].Our results are important for understanding the poten-tial risk to humans from hantavirus infection inMadagascar.We confirm R. rattus as the main reservoir of hanta-

virus in Madagascar. Sequencing results from a subset ofsamples form a distinct cluster with Anjozorobe virus.The detection of two infected M. musculus may suggestspillover infections. Spillover infections in sympatrichosts of closely related species or genera have been

Fig. 2 Geographical distribution of hantavirus detected in small terrestrial mammals trapped in Madagascar

Raharinosy et al. Virology Journal (2018) 15:83 Page 6 of 11

reported before for Dobrava orthohantavirus (DOBV)[49] and Tula orthohantavirus (TULV) [11]. Our resultsare consistent with the idea of that whilst hantaviruseshave preferred host species, some spillover infectionscan occur [11]. However, the importance of such sec-ondary spillover hosts for the persistence and dynamicsof infection is unclear. For TULV, the spatial distribution

of infected individuals from secondary host species andtheir apparent clearance of the virus suggest they play aminor role, if any [11].We did not detect any hantavirus in two other intro-

duced mammals, R. norvegicus and S. murinus. R. norve-gicus is the main reservoir of Seoul orthohantavirus, avirus that is found predominantly in the Old World

Table 2 Estimated parameters from individual-levels variables of Rattus rattus based on Generalized Linear Mixed Models

Variable Models OR CI 95% AICa LRT p_value

Weight Linear 2.43 1.84–3.23 575 41.72 < 0.001

Non-Linear 0.96 0.78–1.18 576.8 0.15 0.7

Sex a 0.75 0.47–1.22 575.6 1.36 0.25

Habitat a House 1 573.28 11.68 0.04

Outside house 2.87 0.87–9.49

Abattoir 0.97 0.22–4.28

Market 2.32 0.94–5.72

Exterior line-trap 1.89 0.93–3.84

Low-ground 2.68 1.36–5.28

Habitat a House 1 567.9 9.04 0.003

Exterior 2.19 1.3–3.69

Variable Habitat: where animal was caught in house and exterior including outside house, Abattoir, market, Exterior line-trap, area around rice fields or other lowground with vegetation. Variables in bold had significant effect on hantavirus infection with p_value< 0.05OR Odds Ratio, CI Confidence interval, AIC Akaike information criterionaAll models included a linear effect of weight due to convergence issues (see text)

Fig. 3 Correlation between the first two principal components and the six climatic variables. Bio 1: Annual Mean Temperature, Bio 2: MeanDiurnal Range in temperature, Bio 4: Temperature Seasonality, Bio12: Annual Precipitation, Bio 15: Precipitation Seasonality, Bio 17: Precipitation ofDriest Quarter. Axis x or F1 and Axis y or F2 represented 66 and 22% of the climatic data information respectively. F1 was positively correlatedwith Bio 12 and Bio 17. However, F2 was negatively correlated with Bio 2 and Bio 15

Raharinosy et al. Virology Journal (2018) 15:83 Page 7 of 11

[50–52], but thought to have a more widespread global dis-tribution [2] and recently confirmed using molecular tech-niques as causing human infections in Europe [53, 54]. S.murinus has been found infected with Thottapalayamorthohantavirus in India, Thailand and Indonesia. As oursample size for R. norvegicus and S. murinus were relativelysmall (n = 124 and n = 76 respectively), we cannot excludethe possibility that they are infected with hantavirus inMadagascar.The global prevalence of infection amongst R. rattus

was 12.4% (111/897), which is similar to the prevalence

of Mayotte virus observed in R. rattus populations onMayotte Island (18.1%: 29/160) [26]. However, we de-tected variation in prevalence at two spatial scales.Large-scale spatial variation was not driven by any rural-urban split but was related to climate, whilst there wasalso within-site variation related to habitat type. Spatialprevalence patterns at both scales may be associatedwith variation in transmission rates due to changingcontact rates between potential hosts and infective parti-cles. Serological infection may have found more rats thathave been exposed to hantaviruses, but using a more

Table 3 Effect of site-level independent variables of Rattus rattus hantavirus infection: comparison of linear and non-linear GeneralizedLinear Mixed Models

Variable Model OR CI 95% LRT DF p-value AIC

No additional variables (weight + habitat) 567.9

Bioclimat Subarid 1 5.5 3 0.14 568.4

Dry 7.15 1–51.26

Humid 8.34 1.22–56.83

Subhumid 6.24 1–38.79

Season 1.24 0.39–3.88 0.13 1 0.72 569.8

Rural/Urban 1.03 0.35–2.97 0.002 1 0.96 570.2

Rattus rattus (Rr) abundance Rr caught within houses Linear 0.83 0.51–1.36 0.53 1 0.47 569.4

Non-linear 1.22 0.78–1.93 0.72 1 0.4 570.7

Rr caught outside trap lines Linear 1.18 0.7–1.99 0.39 1 0.53 569.5

Non-linear 1.08 0.62–1.9 0.08 1 0.73 571.5

Shannon diversity index Linear 1 0.63–1.59 0.00 1 1 569.9

Non-linear 1.08 0.62–1.9 0.21 1 0.65 571.7

Evenness diversity index Linear 0.84 0.49–1.41 0.45 1 0.5 569.5

Non-linear 0.92 0.58–1.47 0.12 1 0.73 571.45

Mean annual temperature Linear 1.11 0.66–1.87 0.15 1 0.7 569.8

Non-linear 1.09 0.61–1.95 0.09 1 0.76 571.7

Mean diurnal range in temperature Linear 0.56 0.33–0.96 4.53 1 0.03 565.4

Non-linear 0.58 0.31–1.09 3.19 1 0.07 564.2

Temperature seasonality Linear 0.6 0.36–1.01 3.67 1 0.06 566.3

Non-linear 0.6 0.28–1.26 1.79 1 0.18 566.5

Annual precipitation Linear 2 1.28–3.12 8.67 1 0.003 561.3

Non-linear 0.75 0.54–1.04 3.22 1 0.07 560.0

Precipitation seasonality Linear 0.8 0.47–1.36 0.67 1 0.41 569.3

Non-linear 1.02 0.53–1.97 0.004 1 0.95 571.3

Precipitation of driest quarter Linear 1.45 0.89–2.37 2.16 1 0.14 567.8

Non-linear 1.05 0.49–2.23 0.02 1 0.9 569.7

PCA Axis 1 (F1) Linear 1.31 1.02–1.68 4.36 1 0.04 565.6

Non-linear 0.82 0.69–0.98 5.21 1 0.02 562.4

PCA Axis 2 (F2) Linear 1.08 0.71–1.66 0.13 1 0.72 569.8

Non-linear 1.16 0.8–1.68 0.61 1 0.43 571.2

Weight and Habitat (significant variables from Table 2) were included in each model. For non-linear models, likelihood ratio tests (LRT) compare the non-linearmodel with the linear model. For linear models, LRT compare the linear model with the best model from Table 2 (includes weight and two-level habitat variable).P-values < 0.05 are shown in bold; p values > 0.05 and < 0.20 are shown in italics. AIC values within 2 of the model with the lowest AIC are shown in boldOR Odds Ratio, CI Confidence interval, AIC Akaike information criterion, LRT likelihood ratio tests, DF Degree of Freedom, PCA Principal Components Analysis

Raharinosy et al. Virology Journal (2018) 15:83 Page 8 of 11

conservative test that detects current presence of DNAis arguably more relevant for the potential infection riskfor humans as these individuals are actively infected,moreover uncertainty about specificity of serologicaltesting and the potential for cross-reaction with anti-bodies against other viruses could have resulted in noisewithin the dataset (false positives). Furthermore, RT-PCR could allow the detection of new variant of ANJOVor a new orthohantavirus which was of importance forus since to date, hantavirus described in Madagascar,were detected from R. rattus collected in a small regionof Madagascar [25].At the national scale, sites in humid environments pre-

sented the highest rates of hantavirus infection. Climaticdrivers, including increased precipitation, have beenlinked to inter-annual variation of hantavirus cases inhumans in China and the Four Corners Region of theUSA [33, 34, 55]. Various mechanisms for these relation-ships have been proposed, including increased vegetationgrowth leading to more abundant rodent populationsand increased contact and viral transmission betweenrodents and between rodents and humans [34, 55]. Apositive relationship between host abundance and infec-tion prevalence has been found for Puumala orthohanta-virus in a spatiotemporal study of bank vole populations[39]. However, some other studies have found only weakrelationships between principal host abundance andprevalence [56], and in our study we did not observe anysignificant association between rat abundance andhantavirus prevalence. The absence of strong effects ofhost abundance may reflect difficulties in measuringabundance in a relevant way, for example current preva-lence may depend on previous host abundance ratherthan current host abundance, or because the relationshipbetween abundance and contact rates is not linear [57].Alternatively, other factors may drive the association be-tween high precipitation and high prevalence throughimpacts on contact rates.Rather than absolute rat abundance, the positive associ-

ation may instead be linked to variation in the seasonalityof reproduction. As indicated by the PCA analysis, siteswith high precipitation are also characterized by less sea-sonality. Rats inhabiting such environments, where thereare sufficient food resources throughout the year to allowreproduction, may exhibit behaviors that are linked tohantavirus transmission throughout the year (e.g. aggres-sive behaviors or behaviors associated with maintenanceof territories). As humidity can maintain the infectivityand the stability of virus in the ex-vivo environment [58],an alternative explanation could be that increased contactbetween rodents and infective particles increases hanta-virus infection rates in humid areas.Within sites, rats trapped outside were more likely to be

infected than rats trapped inside houses. This variation

could also be due to differences in rat contact structuresand virus persistence. Contact structures between rats liv-ing in houses are likely to be very different from rats inha-biting outdoors. If rats typically inhabit a single house,each house may represent a single small and relatively iso-lated population with a much-reduced contact structure.A non-exclusive alternative hypothesis could be that per-sistence of hantavirus in the environment is reducedwithin houses.Several studies of hantavirus in reservoir populations

have found an apparent dilution effect, with lower preva-lence in sites with increased mammal diversity [37–39].Such an effect can occur through a variety of mecha-nisms such as decreases in host density due to competi-tion or reductions in encounter rates between infectedand susceptible hosts. In studies of Sin Nombre virusthere has been some evidence that increased mammaldiversity leads to reduced intraspecific contact rates inthe principal host species [37]. In our study we found noevidence of an effect of diversity, measured at the site-level, on hantavirus prevalence and it seems unlikely thatcommunity diversity explains the difference in infectionprobability between inside and outside rats, as all thespecies trapped were caught in both types of locations(apart from M. musculus, which could only be capturedby Sherman traps which were restricted to houses, andthe rarely trapped H. semispinosus). The lack of a dilu-tion effect in our study may suggest that whilst an effectof community composition appears to be common forhantaviruses in reservoir populations inhabiting naturaland semi-natural habitats located in several parts of theworld [59], it is not a feature of hantaviruses in a perido-mestic context. Alternatively, as for abundance, ourmeasures of community diversity may be inappropriateto capture the underlying mechanism. For example, theeffects of community composition on Puumala ortho-hantavirus were driven by abundance of a specific spe-cies, rather than overall community diversity [39, 60].Thus, more detailed analyses, as well as data from morediverse sites, would further improve our understandingof the potential mechanisms behind the spatial variationin prevalence.

ConclusionsTo conclude, we report the widespread distribution ofTHAIV in the peridomestic rat, R. rattus, captured inboth urban and rural sites of Madagascar, with highestprevalence in humid areas of the island. Thus, althoughthe potential risk of transmission to humans may also bewidespread, our results provide a first indication of areaswhere the risk may be higher. The reduced infectionprobabilities in rats living in houses may decrease therisk of transmission to humans, but data on human ex-posure is necessary to properly evaluate the risk.

Raharinosy et al. Virology Journal (2018) 15:83 Page 9 of 11

AcknowledgementsWe thank those who facilitated the survey: householders, heads of fokontany,local administration and health authorities from Ministry of Health. We wouldlike to express our gratitude to the staff of the Plague Central Laboratory Unit,Institut Pasteur de Madagascar: Dr. Minoarisoa Rajerison who facilitated thisstudy; Corinne Rahaingosoamamitiana and Soanandrasana Rahelinirina forhelping to conduct and organize the field work. We would also like to thank Dr.Fanjasoa Rakotomanana and Dr. Lalaina Arivony Nomenjanahary assistance inthe field trips and technical and field support.

FundingThis work was supported by the Institut Pasteur de Madagascar (InternalProject through ZORA: Zoonoses, Rodent and Arboviruses) and WellcomeTrust Fellowships to ST (#081705, #095171). VR was also supported thoughGirard’s fellowship undergraduate program from the Institut Pasteur deMadagascar. The funders had no role in study design, data collection andanalysis, decision to publish, or preparation of the manuscript.

Availability of data and materialsThe datasets used and/or analysed during the current study are availablefrom the corresponding author upon reasonable request.

Authors’ contributionsVR performed the viral diagnosis and drafted the manuscript. VR and STanalyzed the statistical data. MMO, FMA, SAF, JPR and SA coordinated thefieldwork and participated to the sampling collection. MMO, SFA, ST andJMH designed the field study. CF helped in the phylogenetic analyses andrevised the paper. JMH, ST and DADR coordinated the study and revised thepaper. All authors reviewed the paper. All authors read and approved thefinal manuscript.

Ethics approvalThis study was conducted according to the institutional ethical guidelines.The Institut Pasteur de Madagascar is guided by the International GuidingPrinciples for Biomedical Research Involving Animals. This Institution complywith the following laws, regulations, and policies governing the care and useof laboratory animals (Directive 2010/63/EU revising Directive 86/609/EEC onthe protection of animals used for scientific purposes was adopted on 22September 2010; National charter related to Ethics on animalexperimentation, Ministry of Education and Research, Ministry of agriculture,livestock and fisheries; 3 April 2008; Charter of the Institut Pasteur related toEthics on animal experimentation). Moreover, this institution is approved bythe US Office of Laboratory Animal Welfare (Animal Welfare Assurancenumber F17–00356).

Competing interestsThe authors declare that they have no competing interests.

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Author details1Virology Unit, Institute Pasteur de Madagascar, Ambatofotsikely, BP 1274Antananarivo, Madagascar. 2Ecole Doctorale des Sciences de la Vie et del’Environnement, Equipe Pathogènes et Diversité Moléculaire, Faculté desSciences, Université d’Antananarivo, Antananarivo, Madagascar. 3Plague Unit,Institut Pasteur de Madagascar, Antananarivo, Madagascar. 4Département deBiochimie Fondamentale et Appliquée, Faculté des Sciences, Universitéd’Antananarivo, Antananarivo, Madagascar. 5Institute of Biological andEnvironmental Sciences, University of Aberdeen, Aberdeen, UK.

Received: 19 January 2018 Accepted: 30 April 2018

References1. Yanagihara R, Gu SH, Arai S, Kang HJ, Song JW. Hantaviruses: rediscovery

and new beginnings. Virus Res. 2014;187:6–14.2. Jonsson CB, Figueiredo LT, Vapalahti O. A global perspective on hantavirus

ecology, epidemiology, and disease. Clin Microbiol Rev. 2010;23:412–41.

3. Muranyi W, Bahr U, Zeier M, van der Woude FJ. Hantavirus infection. J AmSoc Nephrol. 2005;16:3669–79.

4. Krautkramer E, Zeier M, Plyusnin A. Hantavirus infection: an emerginginfectious disease causing acute renal failure. Kidney Int. 2013;83:23–7.

5. Macneil A, Nichol ST, Spiropoulou CF. Hantavirus pulmonary syndrome.Virus Res. 2011;162:138–47.

6. Dearing MD, Dizney L. Ecology of hantavirus in a changing world. Ann N YAcad Sci. 2010;1195:99–112.

7. Hu T, Fan Q, Hu X, Deng B, Chen G, Gu L, Li M, Zheng Y, Yuan G, Qiu W, etal. Molecular and serological evidence for Seoul virus in rats (Rattusnorvegicus) in Zhangmu, Tibet, China. Arch Virol. 2015;160:1353–7.

8. Vapalahti O, Kallio-Kokko H, Salonen EM, Brummer-Korvenkontio M, VaheriA. Cloning and sequencing of Puumala virus Sotkamo strain S and M RNAsegments: evidence for strain variation in hantaviruses and expression ofthe nucleocapsid protein. J Gen Virol. 1992;73(Pt 4):829–38.

9. Mills JN, Schmidt K, Ellis BA, Calderon G, Enria DA, Ksiazek TG. A longitudinalstudy of hantavirus infection in three sympatric reservoir species inagroecosystems on the argentine pampa. Vector Borne Zoonotic Dis. 2007;7:229–40.

10. Schlegel M, Radosa L, Rosenfeld UM, Schmidt S, Triebenbacher C, Lohr PW,Fuchs D, Heroldova M, Janova E, Stanko M, et al. Broad geographicaldistribution and high genetic diversity of shrew-borne Seewis hantavirus inCentral Europe. Virus Genes. 2012;45:48–55.

11. Schmidt S, Saxenhofer M, Drewes S, Schlegel M, Wanka KM, Frank R, KlimpelS, von Blanckenhagen F, Maaz D, Herden C, et al. High genetic structuringof Tula hantavirus. Arch Virol. 2016;161:1135–49.

12. Botten J, Mirowsky K, Ye C, Gottlieb K, Saavedra M, Ponce L, Hjelle B.Shedding and intracage transmission of sin Nombre hantavirus in the deermouse (Peromyscus maniculatus) model. J Virol. 2002;76:7587–94.

13. Hardestam J, Karlsson M, Falk KI, Olsson G, Klingstrom J, Lundkvist A.Puumala hantavirus excretion kinetics in bank voles (Myodes glareolus).Emerg Infect Dis. 2008;14:1209–15.

14. Bouloy M, Zeller H. Hantavirus: recent data and review of the literature. BullSoc Pathol Exot. 2000;93:177–80.

15. Padula PJ, Edelstein A, Miguel SD, Lopez NM, Rossi CM, Rabinovich RD.Hantavirus pulmonary syndrome outbreak in Argentina: molecular evidencefor person-to-person transmission of Andes virus. Virology. 1998;241:323–30.

16. Martinez VP, Bellomo C, San Juan J, Pinna D, Forlenza R, Elder M, Padula PJ.Person-to-person transmission of Andes virus. Emerg Infect Dis. 2005;11:1848–53.

17. Klempa B, Fichet-Calvet E, Lecompte E, Auste B, Aniskin V, Meisel H, BarriereP, Koivogui L, ter Meulen J, Kruger DH. Novel hantavirus sequences inshrew, Guinea. Emerg Infect Dis. 2007;13:520–2.

18. Klempa B, Fichet-Calvet E, Lecompte E, Auste B, Aniskin V, Meisel H, DenysC, Koivogui L, ter Meulen J, Kruger DH. Hantavirus in African wood mouse,Guinea. Emerg Infect Dis. 2006;12:838–40.

19. Kang HJ, Kadjo B, Dubey S, Jacquet F, Yanagihara R. Molecular evolution ofAzagny virus, a newfound hantavirus harbored by the west African pygmyshrew (Crocidura obscurior) in cote d'Ivoire. Virol J. 2011;8:373.

20. Sumibcay L, Kadjo B, Gu SH, Kang HJ, Lim BK, Cook JA, Song JW, YanagiharaR. Divergent lineage of a novel hantavirus in the banana pipistrelle(Neoromicia nanus) in cote d’Ivoire. Virol J. 2012;9:34.

21. Meheretu Y, Cizkova D, Tesikova J, Welegerima K, Tomas Z, Kidane D,Girmay K, Schmidt-Chanasit J, Bryja J, Gunther S, et al. High diversity of RNAviruses in rodents, Ethiopia. Emerg Infect Dis. 2012;18:2047–50.

22. Gu SH, Nicolas V, Lalis A, Sathirapongsasuti N, Yanagihara R. Completegenome sequence and molecular phylogeny of a newfound hantavirusharbored by the Doucet's musk shrew (Crocidura douceti) in Guinea. InfectGenet Evol. 2013;20:118–23.

23. Witkowski PT, Drexler JF, Kallies R, Lickova M, Bokorova S, Mananga GD,Szemes T, Leroy EM, Kruger DH, Drosten C, Klempa B. Phylogenetic analysisof a newfound bat-borne hantavirus supports a laurasiatherian hostassociation for ancestral mammalian hantaviruses. Infect Genet Evol. 2016;41:113–9.

24. Weiss S, Witkowski PT, Auste B, Nowak K, Weber N, Fahr J, Mombouli JV,Wolfe ND, Drexler JF, Drosten C, et al. Hantavirus in bat, Sierra Leone. EmergInfect Dis. 2012;18:159–61.

25. Reynes JM, Razafindralambo NK, Lacoste V, Olive MM, Barivelo TA,Soarimalala V, Heraud JM, Lavergne A. Anjozorobe hantavirus, a newgenetic variant of Thailand virus detected in rodents from Madagascar.Vector Borne Zoonotic Dis. 2014;14:212–9.

Raharinosy et al. Virology Journal (2018) 15:83 Page 10 of 11

26. Filippone C, Castel G, Murri S, Beaulieux F, Ermonval M, Jallet C, Wise EL, EllisRJ, Marston DA, McElhinney LM, et al. Discovery of hantavirus circulatingamong Rattus rattus in French Mayotte island, Indian Ocean. J Gen Virol.2016;97:1060–5.

27. Elwell MR, Ward GS, Tingpalapong M, LeDuc JW. Serologic evidence ofHantaan-like virus in rodents and man in Thailand. Southeast Asian J TropMed Public Health. 1985;16:349–54.

28. Nakounne E, Selekon B, Morvan J. Microbiological surveillance: viralhemorrhagic fever in Central African Republic: current serological data inman. Bull Soc Pathol Exot. 2000;93:340–7.

29. Klempa B, Koivogui L, Sylla O, Koulemou K, Auste B, Kruger DH, ter MeulenJ. Serological evidence of human hantavirus infections in Guinea, WestAfrica. J Infect Dis. 2010;201:1031–4.

30. Witkowski PT, Leendertz SA, Auste B, Akoua-Koffi C, Schubert G, Klempa B,Muyembe-Tamfum JJ, Karhemere S, Leendertz FH, Kruger DH. Humanseroprevalence indicating hantavirus infections in tropical rainforests of coted'Ivoire and Democratic Republic of Congo. Front Microbiol. 2015;6:518.

31. Heinemann P, Tia M, Alabi A, Anon JC, Auste B, Essbauer S, Gnionsahe A,Kigninlman H, Klempa B, Kraef C, et al. Human infections by non-rodent-associated hantaviruses in Africa. J Infect Dis. 2016;214:1507–11.

32. Gamage CD, Yasuda SP, Nishio S, Kularatne SA, Weerakoon K, Rajapakse J,Nwafor-Okoli C, Lee RB, Obayashi Y, Yoshimatsu K, et al. Serologicalevidence of Thailand virus-related hantavirus infection among suspectedleptospirosis patients in Kandy, Sri Lanka. Jpn J Infect Dis. 2011;64:72–5.

33. Fang LQ, Wang XJ, Liang S, Li YL, Song SX, Zhang WY, Qian Q, Li YP, Wei L,Wang ZQ, et al. Spatiotemporal trends and climatic factors of hemorrhagicfever with renal syndrome epidemic in Shandong Province, China. PLoSNegl Trop Dis. 2010;4:e789.

34. Engelthaler DM, Mosley DG, Cheek JE, Levy CE, Komatsu KK, Ettestad P,Davis T, Tanda DT, Miller L, Frampton JW, et al. Climatic and environmentalpatterns associated with hantavirus pulmonary syndrome, four cornersregion, United States. Emerg Infect Dis. 1999;5:87–94.

35. Olsson GE, Ahlm C, Elgh F, Verlemyr AC, White N, Juto P, Palo RT. Hantavirusantibody occurrence in bank voles (Clethrionomys glareolus) during a volepopulation cycle. J Wildl Dis. 2003;39:299–305.

36. Yates TL, Mills JN, Parmenter CA, Ksiazek TG, Parmenter RR, Vande Castle JR,Calisher CH, Nichol ST, Abbott KD, Young JC. The ecology and evolutionaryhistory of an emergent disease: hantavirus pulmonary syndrome: evidencefrom two El Niño episodes in the American southwest suggests that ElNiño–driven precipitation, the initial catalyst of a trophic cascade thatresults in a delayed density-dependent rodent response, is sufficient topredict heightened risk for human contraction of hantavirus pulmonarysyndrome. Bioscience. 2002;52:989–98.

37. Clay CA, Lehmer EM, St Jeor S, Dearing MD. Testing mechanisms of thedilution effect: deer mice encounter rates, sin Nombre virus prevalence andspecies diversity. EcoHealth. 2009;6:250–9.

38. Carver S, Kuenzi A, Bagamian KH, Mills JN, Rollin PE, Zanto SN, Douglass R. Atemporal dilution effect: hantavirus infection in deer mice and theintermittent presence of voles in Montana. Oecologia. 2011;166:713–21.

39. Voutilainen L, Savola S, Kallio ER, Laakkonen J, Vaheri A, Vapalahti O,Henttonen H. Environmental change and disease dynamics: effects ofintensive forest management on Puumala hantavirus infection in borealbank vole populations. PLoS One. 2012;7:e39452.

40. Keesing F, Holt RD, Ostfeld RS. Effects of species diversity on disease risk.Ecol Lett. 2006;9:485–98.

41. Burnham KP, Anderson DR. Model selection and multimodel inference NewYork. New York: Springer; 2002.

42. Pielou EC. The measurement of diversity in different types of biologicalcollections. J Theoret Biol. 1966;13:131–44.

43. Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effectsmodels using lme4. J Stat Softw. 2015;67:48. https://doi.org/10.18637/jss.v18067.i18601.

44. Chessel D, Dufour AB, Thioulouse J. The ade4 package-I-one-table methods.R news. 2004;4:5–10. http://pbil.univ-lyon1.fr/JTHome/ref/ade4-Rnews.pdf.Accessed 10 Oct 2017

45. Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P. Vegan: communityecology package. R package version 2. 4–2. 2017. https://CRAN.R-project.org/package=vegan. Accessed 10 Oct 2017.

46. R Development Core Team. R: a language and environment for statisticalcomputing. Vienna: R Foundation for Statistical Computing; 2016. https://www.R-project.org/ Accessed 10 Oct 2017

47. Heroldova M, Pejcoch M, Bryja J, Janova E, Suchomel J, Tkadlec E. Tula virusin populations of small terrestrial mammals in a rural landscape. VectorBorne Zoonotic Dis. 2010;10:599–603.

48. Dobly A, Yzoard C, Cochez C, Ducoffre G, Aerts M, Roels S, Heyman P.Spatiotemporal dynamics of Puumala hantavirus in suburban reservoirrodent populations. J Vector Ecol. 2012;37:276–83.

49. Weidmann M, Schmidt P, Vackova M, Krivanec K, Munclinger P, Hufert FT.Identification of genetic evidence for dobrava virus spillover in rodents bynested reverse transcription (RT)-PCR and TaqMan RT-PCR. J Clin Microbiol.2005;43:808–12.

50. Zhang YZ, Zhang FX, Wang JB, Zhao ZW, Li MH, Chen HX, Zou Y, PlyusninA. Hantaviruses in rodents and humans, Inner Mongolia AutonomousRegion, China. Emerg Infect Dis. 2009;15:885–91.

51. Zhang YZ, Lin XD, Shi NF, Wang W, Liao XW, Guo WP, Fan FN, Huang XM, LiMH, Li MF, et al. Hantaviruses in small mammals and humans in the coastalregion of Zhejiang Province, China. J Med Virol. 2010;82:987–95.

52. Lin XD, Guo WP, Wang W, Zou Y, Hao ZY, Zhou DJ, Dong X, Qu YG, Li MH,Tian HF, et al. Migration of Norway rats resulted in the worldwidedistribution of Seoul hantavirus today. J Virol. 2012;86:972–81.

53. Reynes JM, Carli D, Bour JB, Boudjeltia S, Dewilde A, Gerbier G, NussbaumerT, Jacomo V, Rapt MP, Rollin PE, Septfons A. Seoul virus infection in humans,France, 2014-2016. Emerg Infect Dis. 2017;23:973–7.

54. Jameson LJ, Logue CH, Atkinson B, Baker N, Galbraith SE, Carroll MW, BrooksT, Hewson R. The continued emergence of hantaviruses: isolation of a Seoulvirus implicated in human disease, United Kingdom, October 2012. EuroSurveill. 2013;18:4–7.

55. Hjelle B, Glass GE. Outbreak of hantavirus infection in the four cornersregion of the United States in the wake of the 1997-1998 El Nino-southernoscillation. J Infect Dis. 2000(181):1569–73.

56. Clay CA, Lehmer EM, Jeor SS, Dearing MD. Sin nombre virus and rodentspecies diversity: a test of the dilution and amplification hypotheses. PLoSOne. 2009;4:e6467.

57. Smith MJ, Telfer S, Kallio ER, Burthe S, Cook AR, Lambin X, Begon M. Host-pathogen time series data in wildlife support a transmission functionbetween density and frequency dependence. Proc Natl Acad Sci U S A.2009;106:7905–9.

58. Hardestam J, Simon M, Hedlund KO, Vaheri A, Klingstrom J, Lundkvist A. Exvivo stability of the rodent-borne Hantaan virus in comparison to that ofarthropod-borne members of the Bunyaviridae family. Appl EnvironMicrobiol. 2007;73:2547–51.

59. Dearing MD, Clay C, Lehmer E, Dizney L. The roles of community diversityand contact rates on pathogen prevalence. J Mammal. 2015;96:29–36.

60. Tersago K, Schreurs A, Linard C, Verhagen R, Van Dongen S, Leirs H.Population, environmental, and community effects on local bank vole(Myodes glareolus) Puumala virus infection in an area with low humanincidence. Vector Borne Zoonotic Dis. 2008;8:235–44.

Raharinosy et al. Virology Journal (2018) 15:83 Page 11 of 11